GSI Technology rolls out servers for facial recognition and re-identification | Biometric Update
GSI Technology introduces Leda-E and Leda-S servers, which utilize proprietary compute-in-memory architecture to drastically enhance processing speed for massive datasets. By eliminating input-output bottlenecks, these systems enable parallel processing for complex tasks like facial recognition and object detection, offering a significant performance leap over traditional sequential processors. The technology delivers superior power efficiency, consuming notably less energy than conventional CPU-only systems while supporting advanced features like zero-shot learning. This efficiency is critical for edge computing applications, allowing for high-throughput biometric processing without the excessive thermal and energy costs typically associated with high-performance computing at the edge. This innovation is relevant to open data because it enables the rapid, cost-effective analysis of large-scale public datasets, such as biometric archives. By reducing the computational barriers to processing massive volumes of information, it facilitates more accessible and efficient open data initiatives in security and identification fields.
Source: biometricupdate.comPublished on 2024-04-05
Related news
- Response to NTIA Request for Comment: “Dual Use Foundation Artificial Intelligence Models with Widely Available Model Weights”
- El 85% de los panameños considera que ‘no es fácil’ encontrar empleo en Panamá
- La deforestación mundial se mantiene alta, aunque con mejoras en Brasil y Colombia
- A Process Formality